views.py 9.45 KB
Newer Older
1 2 3 4
#!/usr/bin/env python3
"""
    Routes
"""
5
import copy
6 7 8
import json
import os
from collections import OrderedDict
9
from io import BytesIO
10 11
from pathlib import Path

12
from commonmark import commonmark
13
from flask import Flask, jsonify, redirect, render_template, request, url_for
14

15 16 17 18 19 20 21
import tabulator
from validata_ui import app
from validata_ui.util import (ValidataSource, flash_error, flash_info,
                              flash_success, flash_warning)
from validata_ui.validate_helper import ValidatorHelper
from validata_validate import csv_helpers
from validata_validate.loaders import custom_loaders
22

23 24

def extract_source_data(source: ValidataSource, preview_rows_nb=5):
25
    """ Computes table preview """
26 27 28

    def stringify(val):
        """ Transform value into string """
Pierre Dittgen's avatar
Pierre Dittgen committed
29
        return '' if val is None else str(val)
30

31 32
    header = None
    rows = []
Pierre Dittgen's avatar
Pierre Dittgen committed
33
    nb_rows = 0
34

35
    options = {}
36
    if source.format == "csv":
37 38
        options['delimiter'] = csv_helpers.detect_dialect(source.data, format=source.format, scheme=source.scheme,
                                                          custom_loaders=custom_loaders).delimiter
39
    with tabulator.Stream(source.data, format=source.format, scheme=source.scheme, custom_loaders=custom_loaders,
40
                          **options) as stream:
41 42
        for row in stream:
            if header is None:
43
                header = ['' if v is None else v for v in row]
44
            else:
45
                rows.append(list(map(stringify, row)))
Pierre Dittgen's avatar
Pierre Dittgen committed
46
                nb_rows += 1
47
    preview_rows_nb = min(preview_rows_nb, nb_rows)
48 49
    return {'header': header,
            'rows_nb': nb_rows,
50 51 52
            'data_rows': rows,
            'preview_rows_nb': preview_rows_nb,
            'preview_rows': rows[:preview_rows_nb]}
53 54


55 56 57 58 59 60 61 62 63 64
def improve_errors(errors):
    """ add context to errors, converts markdown content to HTML"""

    def improve_err(err):
        """ Adds context info based on row-nb presence and converts content to HTML"""

        # Context
        update_keys = {
            'context': 'body' if 'row-number' in err and not err['row-number'] is None else 'table',
        }
Pierre Dittgen's avatar
Pierre Dittgen committed
65

66 67 68 69 70 71 72 73 74
        # markdown to HTML (with default values for 'title' and 'content')
        # Use default values to insure right error display in validation report
        # until validata.validate finished its message migration
        if not 'title' in err:
            update_keys['title'] = '[{}]'.format(err['code'])
        if not 'message' in err or err['message'] is None:
            update_keys['message'] = '[{}]'.format(err['code'])
        md_content = '*content soon available*' if not 'content' in err else err['content']
        update_keys['content'] = commonmark(md_content)
75

76 77 78
        return {**err, **update_keys}

    return list(map(improve_err, errors))
79 80


81
def create_validata_report(goodtables_report, schema):
82 83 84 85 86 87
    """ Creates an error report easier to handle and display in templates:
        - only one table
        - errors are contextualized
        - error-counts is ok
        - errors are grouped by lines
        - errors are separated into "structure" and "body"
88
        - error messages are improved
89 90 91 92 93 94 95 96 97 98 99
    """
    report = copy.deepcopy(goodtables_report)

    # One table is enough
    del report['table-count']
    report['table'] = report['tables'][0]
    del report['tables']
    del report['table']['error-count']
    del report['table']['time']
    del report['table']['valid']
    del report['valid']
100 101 102
    # use _ instead of - to ease information picking in jinja2 template
    report['table']['row_count'] = report['table']['row-count']

Pierre Dittgen's avatar
Pierre Dittgen committed
103 104 105 106
    # Handy col_count info
    headers = report['table'].get('headers', [])
    report['table']['col_count'] = len(headers)

Pierre Dittgen's avatar
Pierre Dittgen committed
107
    # Computes column info
108
    schema_fields = schema.get('fields', [])
Pierre Dittgen's avatar
Pierre Dittgen committed
109 110 111 112
    fields_dict = {f['name']: (f.get('title', 'titre non défini'), f.get('description', '')) for f in schema_fields}
    report['table']['headers_title'] = [fields_dict[h][0] if h in fields_dict else 'colonne inconnue' for h in headers]
    report['table']['headers_description'] = [fields_dict[h][1]
                                              if h in fields_dict else 'Cette colonne n\'est pas définie dans le schema' for h in headers]
113

114 115
    # Contextualize errors and convert content from markdown to HTML
    errors = improve_errors(report['table']['errors'])
116 117 118 119 120 121 122 123 124 125 126 127 128 129
    del report['table']['errors']

    # Count errors
    report['error_count'] = len(errors)
    del report['error-count']

    # Then group them in 2 groups : structure and body
    report['table']['errors'] = {'structure': [], 'body': []}
    for err in errors:
        if err['context'] != 'body':
            report['table']['errors']['structure'].append(err)
        else:
            report['table']['errors']['body'].append(err)

Pierre Dittgen's avatar
Pierre Dittgen committed
130
    # Checks if there are structure errors different to invalid-column-delimiter
131 132 133
    report['table']['display_body_errors'] = all(err['code'] == 'invalid-column-delimiter'
                                                 for err in report['table']['errors']['structure'])

134 135 136
    # Remember number of body errors
    report['table']['errors']['body_nb_errors'] = len(report['table']['errors']['body'])

Pierre Dittgen's avatar
Pierre Dittgen committed
137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155
    # Group body errors by row id
    rows = []
    current_row_id = 0
    for err in report['table']['errors']['body']:
        if not 'row-number' in err:
            print('ERR', err)
        row_id = err['row-number']
        del err['row-number']
        del err['context']
        if row_id != current_row_id:
            current_row_id = row_id
            rows.append({'row_id': current_row_id, 'errors': {}})

        column_id = err.get('column-number')
        if column_id is not None:
            del err['column-number']
            rows[-1]['errors'][column_id] = err
        else:
            rows[-1]['errors']['row'] = err
156 157 158
    report['table']['errors']['body_by_rows'] = rows

    del report['table']['errors']['body']
Pierre Dittgen's avatar
Pierre Dittgen committed
159 160 161 162

    return report


163
def validate(schema_code, source: ValidataSource):
164 165
    """ Validate source and display report """

166 167 168 169 170
    try:
        goodtables_report = ValidatorHelper.validate(schema_code, **source.get_goodtables_source())
    except tabulator.exceptions.FormatError:
        flash_error('Erreur : format de fichier non supporté')
        return redirect(url_for('scdl_validator', val_code=schema_code))
171

172 173
    # return jsonify(goodtables_report)

174
    source_data = extract_source_data(source)
175

176
    validata_report = create_validata_report(goodtables_report, ValidatorHelper.schema(schema_code))
177

178
    # return jsonify(validata_report)
Pierre Dittgen's avatar
Pierre Dittgen committed
179

Pierre Dittgen's avatar
Pierre Dittgen committed
180 181
    # Complete report
    val_info = ValidatorHelper.schema_info(schema_code)
182
    return render_template('validation_report.html', title='Rapport de validation',
183
                           val_info=ValidatorHelper.schema_info(schema_code), report=validata_report,
184
                           source=source, source_type=source.type, source_data=source_data,
185
                           report_str=json.dumps(validata_report, sort_keys=True, indent=2),
Pierre Dittgen's avatar
Pierre Dittgen committed
186
                           breadcrumbs=[{'url': url_for('home'), 'title': 'Accueil'},
Pierre Dittgen's avatar
Pierre Dittgen committed
187 188
                                        {'url': url_for('scdl_validator', val_code=schema_code),
                                         'title': val_info['title']}])
189 190


191 192 193 194 195 196 197 198
def bytes_data(f):
    """ Gets bytes data from Werkzeug FileStorage instance """
    iob = BytesIO()
    f.save(iob)
    iob.seek(0)
    return iob.getvalue()


199 200 201 202 203 204
# Routes


@app.route('/')
def home():
    """ Home page """
Pierre Dittgen's avatar
Pierre Dittgen committed
205
    validators = ValidatorHelper.schema_info_list()
Pierre Dittgen's avatar
Pierre Dittgen committed
206
    flash_warning('Ce service est fourni en mode beta - certains problèmes peuvent subsister - nous mettons tout en œuvre pour améliorer son fonctionnement en continu')
207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225
    return render_template('home.html', title='Accueil', validators=validators)


@app.route('/validators')
def validators():
    """ No validators page """
    return redirect(url_for('home'))


@app.route('/validators/<val_code>', methods=['GET', 'POST'])
def scdl_validator(val_code):
    """ Validator page """

    if not ValidatorHelper.schema_exist(val_code):
        flash_error('Validateur [{}] inconnu'.format(val_code))
        return redirect(url_for('home'))

    if request.method == 'GET':

Pierre Dittgen's avatar
Pierre Dittgen committed
226
        val_info = ValidatorHelper.schema_info(val_code)
227 228 229 230
        input_param = request.args.get('input')

        # First form display
        if input_param is None or input_param not in ('url', 'example'):
231
            return render_template('validation_form.html', title=val_info['title'],
232 233 234 235 236 237 238 239 240
                                   val_info=val_info,
                                   breadcrumbs=[{'url': url_for('home'), 'title': 'Accueil'}, ])

        # Process URL
        else:
            url = request.args.get('url')
            if url is None or url == '':
                flash_error("Vous n'avez pas indiqué d'url à valider")
                return redirect(url_for('scdl_validator', val_code=val_code))
241
            return validate(val_code, ValidataSource(url, url, 'url'))
242 243 244 245

    else:  # POST
        input_param = request.form.get('input')
        if input_param is None:
246
            flash_error('Aucun fichier à valider')
247 248 249 250 251 252 253 254
            return redirect(url_for('scdl_validator', val_code=val_code))

        # File validation
        if input_param == 'file':
            f = request.files.get('file')
            if f is None:
                flash_warning("Vous n'avez pas indiqué de fichier à valider")
                return redirect(url_for('scdl_validator', val_code=val_code))
255 256

            return validate(val_code, ValidataSource(bytes_data(f), f.filename, 'file'))
257 258

        return 'Bizarre, vous avez dit bizarre ?'